In general, gasoline and other liquid hydrocarbon fuels boil in the range of about 100.degree. to about 650.degree. F. However, the crude oil from which these fuels are made contains a diverse mixture of hydrocarbons and other compounds which vary widely in molecular weight and therefore boil over a wide range. For example, crude oils are known in which 30 to 60% or more of the total volume of oil is composed of compounds boiling at temperatures above 650.degree. F. Among these are crudes in which about 10% to about 30% or more of the total volume consists of compounds so heavy in molecular weight that they boil above 1025.degree. F. or at least will not boil below 1025.degree. F. at atmospheric pressure.
Because these relatively abundant high boiling components of crude oil are unsuitable for inclusion in gasoline and other liquid hydrocarbon fuels, the petroleum refining industry has developed processes for cracking or breaking the molecules of the high molecular weight, high boiling compounds into smaller molecules which do boil over an appropriate boiling range. The cracking process which is most widely used for this purpose is known as fluid catalytic cracking (FCC). Although the FCC process has reached a highly advanced state, and many modified forms and vari-actions have been developed, their unifying factor is that a vaporized hydrocarbon feedstock is caused to crack at an elevated temperature in contact with a cracking catalyst that is suspended in the feedstock vapors. Upon attainment of the desired degree of molecular weight and boiling point reduction the catalyst is separated from the desired products.
Crude oil in the natural state contains a variety of materials which tend to have quite troublesome effects on FCC processes, and only a portion of these troublesome materials can be economically removed from the crude oil. Among these troublesome materials are coke precursors (such as asphaltenes, polynuclear aromatics, etc.), heavy metals (such as nickel, vanadium, iron, copper, etc.), lighter metals (such as sodium, potassium, etc.), sulfur, nitrogen and others. Certain of these, such as the lighter metals, can be economically removed by desalting operations, which are part of the normal procedure for pretreating crude oil for fluid catalytic cracking. Other materials, such as coke precursors, asphaltenes and the like, tend to break down into coke during the cracking operation, which coke deposits on the catalyst, impairing contact between the hydrocarbon feedstock and the catalyst, and generally reducing its potency or activity level. The heavy metals transfer almost quantitatively from the feedstock to the catalyst surface.
If the catalyst is reused again and again for processing additional feedstock, which is usually the case, the heavy metals can accumulate on the catalyst to the point that they unfavorably alter the composition of the catalyst and/or the nature of its effect upon the feedstock. For example, vanadium tends to form fluxes with certain components of commonly used FCC catalysts, lowering the melting point of portions of the catalyst particles sufficiently so that they begin to sinter and become ineffective cracking catalysts. Accumulations of vanadium and other heavy metals, especially nickel, also "poison" the catalyst. They tend in varying degrees to promote excessive dehydrogenation and aromatic condensation, resulting in excessive production of carbon and gases with consequent impairment of liquid fuel yield. An oil such as a crude or crude fraction or other oil that is particularly abundant in nickel and/or other metals exhibiting similar behavior, while containing relatively large quantities of coke precursors, is referred to herein as a carbo-metallic oil, and represents a particular challenge to the petroleum refiner.
In general, the coke-forming tendency or coke precursor content of an oil can be ascertained by determining the weight percent of carbon remaining after a sample of that oil has been pyrolyzed. The industry accepts this value as a measure of the extent to which a given oil tends to form non-catalytic coke when employed as feedstock in a catalytic cracker. Two established tests are recognized, the Conradson Carbon and Ramsbottom Carbon tests, the former being described in ASTM D189-76 and the latter being described in ASTM Test No. D524-76. In conventional FCC practice, Conradson carbon values on the order of about 0.05 to about 1.0 are regarded as indicative of acceptable feed. The present invention is concerned with the use of hydrocarbon feedstocks which have higher Conradson carbon values and thus exhibit substantially greater potential for coke formation than the usual feeds.
Since the various heavy metals are not of equal catalyst poisoning activity, it is convenient to express the poisoning activity of an oil containing a given poisoning metal or metals in terms of the amount of a single metal which is estimated to have equivalent poisoning activity. Thus, the heavy metals content of an oil can be expressed by the following formula (patterned after that of W. L. Nelson in Oil and Gas Journal, page 143, Oct. 23, 1961) in which the content of each metal present is expressed in parts per million of such metal, as metal, on a weight basis, based on the weight of feed: ##EQU1## According to conventional FCC practice, the heavy metal content of feedstock for FCC-processing is controlled at a relatively low level, e.g., about 0.25 ppm Nickel Equivalents or less. The present invention is concerned with the processing of feedstocks containing metals substantially in excess of this value and which therefore have a significantly greater potential for accumulating on and poisoning catalyst.
The above formula can also be employed as a measure of the accumulation of heavy metals on cracking catalyst, except that the quantity of metal employed in the formula is based on the weight of catalyst (moisture free basis) instead of the weight of feed. In conventional FCC practice, in which a circulating inventory of catalyst is used again and again in the processing of fresh feed, with periodic or continuing minor addition and withdrawal of fresh and spent catalyst, the metal content of the catalyst is maintained at a level which may for example be in the range of about 200 to about 600 ppm Nickel Equivalents. The process of the present invention is concerned with the use of catalyst having a substantially larger metals content, and which therefore has a much greater than normal tendency to promote dehydrogenation, aromatic condensation, gas production or coke formation. Therefore, such higher metals accumulation is normally regarded as quite undesirable in FCC processing.
There has been a long standing interest in the conversion of carbo-metallic oils into gasoline and other liquid fuels. For example, in the 1950s it was suggested that a variety of carbo-metallic oils could be successfully converted to gasoline and other products in the Houdresid process. Turning from the FCC mode of operation, the Houdresid process employed catalyst particles of "granular size" (much larger than conventional FCC catalyst particle size) in a compact gravitating bed, rather than suspending catalyst particles in feed and produce vapors in a fluidized bed.
Although the Houdresid process obviously represented a step forward in dealing with the effects of metal contamination and coke formation on catalyst performance, its productivity was limited. Because its operation was uneconomical, the first Houdresid unit is no longer operating. Thus, for the 25 years which have passed since the Houdresid process was first introduced commercially, the art has continued its arduous search for suitable modifications or alternatives to the FCC process which would permit commercially successful operation on reduced crude and the like. During this period a number of proposals have been made; some have been used commercially to a certain extent.
Several proposals involve treating the heavy oil feed to remove the metal therefrom prior to cracking, such as by hydrotreating, solvent extraction and complexing with Friedel-Crafts catalysts, but those techniques have been criticized as unjustified economically. Another proposal employs a combination cracking process having "dirty oil" and "clean oil" units. Still another proposal blends residual oil with gas oil and controls the quantity of residual oil in the mixture in relation to the equilibrium flash vaporization temperature at the bottom of the riser type cracker unit employed in the process. Still another proposal subjects the feed to a mild preliminary hydrocracking or hydrotreating operation before it is introduced into the cracking unit. It has also been suggested to contact a carbo-metallic oil such as reduced crude with hot taconite pellets to produce gasoline. This is a small sampling of the many proposals which have appeared in the patent literature and technical papers.
Notwithstanding the great effort which has been expended and the fact that each of these proposals overcomes some of the difficulties involved, conventional FCC practice today bears mute testimony to the dearth of carbo-metallic oil-cracking techniques that are both economical and highly practical in terms of technical feasibility. Some crude oils are relatively free of coke precursors or heavy metals or both, and the troublesome components of crude oil are for the most part concentrated in the highest boiling fractions. Accordingly, it has been possible to largely avoid the problems of coke precursors and heavy metals by sacrificing the liquid fuel yield which would be potentially available from the highest boiling fractions. More particularly, conventional FCC practice has employed as feedstock that a fraction of crude oil which boils at about 650.degree. F. to about 1,000.degree. F., such fractions being relatively free of coke precursors and heavy metal contamination. Such feedstock, known as "vacuum gas oil" (VGO) is generally prepared from crude oil by distilling off the fractions boiling below about 650.degree. F. at atmospheric pressure and then separating by further vacuum distillation from the heavier fractions a cut boiling between about 650.degree. F. and about 900.degree. to 1025.degree. F.
The vacuum gas oil is used as feedstock for conventional FCC processing. The heavier fractions are normally employed for a variety of other purposes, such as for instance production of asphalt, residual fuel oil, #6 fuel oil, or marine Bunker C fuel oil, which represents a great waste of the potential value of this portion of the crude oil, especially in light of the great effort and expense which the art has been willing to expend in the attempt to produce generally similar materials from coal and shale oils. The present invention is aimed at the simultaneous cracking of these heavier fractions containing substantial quantities of both coke precursors and heavy metals, and possibly other troublesome components, in conjunction with the lighter oils, thereby increasing the overall yield of gasoline and other hydrocarbon liquid fuels from a given quantity of crude. As indicated above, the present invention by no means constitutes the first attempt to develop such a process, but the long standing recognition of the desirability of cracking carbo-metallic feedstocks, along with the slow progress of the industry toward doing so, show the continuing need for such a process. It is believed that the present process is uniquely advantageous for dealing with the problem of treating such carbo-metallic oils in an economically and technically sound manner.
One method of cracking these high boiling fractions, name Reduced Crude Conversion (RCC) after a particularly common and useful carbo-metallic feed, is disclosed in copending applications Ser. No. 94,092 (now U.S. Pat. No. 4,332,673) and Ser. No. 94,216 (now U.S. Pat. No. 4,341,624), each filed Nov. 14, 1979, for "Carbo-Metallic Oil Conversion" and each being incorporated herein by reference. The oils disclosed as capable of being cracked by the methods of these applications are carbo-metallic oils of which at least about 70 percent boils above 650.degree. F. and which contain a carbon residue on pyrolysis of at least about 1 and at least about 4 parts per million of nickel equivalents of heavy metals. Examples of these oils are crude oils, topped crudes, reduced crudes, residua, and extracts from solvent deasphalting.
The cracking reaction for the method disclosed in application Ser. No. 94,216 (now U.S. Pat. No. 4,341,624) is sufficiently severe to convert 50% or more of the feedstock to gasoline per pass and produce coke in the amount of 6 to 14% by weight based on weight of fresh feed. In a typical RCC cracking process the ratio of weight of catalyst to weight of feedstock is from about 3 to about 18, coke is laid down on the catalyst in amounts in the range of about 0.3 to about 3 percent by weight based on the weight of the catalyst, and heavy metals accumulate on the catalyst to a concentration of from about 3000 to about 30,000 ppm nickel equivalents.
For purposes of this application, the term "heavy metals" refers to nickel, vanadium, copper and iron, although trace amounts of other heavy metal elements may sometimes be present. The total amount of heavy metals in a carbo-metallic oil feed is usually comprised principally of nickel and vanadium (more than 90 weight percent based on weight of total heavy metals). Although the feed may contain principally either nickel or vanadium, vanadium is often present relative to nickel at weight ratios of 1.0 or greater. During the cracking process, the heavy metal inventory of the feed transfers almost quantitatively from the feedstock oil to the catalyst particles. Although the heavy metals initially tend to deposit near the surface of a porous catalyst matrix, significant proportions of these metals migrate to the interior of the catalyst matrix where they can deposit on zeolite or other catalytic promoters carried within the matrix.
Heavy metal deposits on the catalyst, particularly nickel, have in the past catalyzed undesirable dehydrogenation reactions which increased coke production and decreased the selectivity of the catalyst for gasoline production. Although these dehydrogenation reactions produce hydrogen gas in the exiting product stream, it appears that this hydrogen does not significantly increase within the riser the hydrogen to carbon ratio of the hydrogen deficient hydrocarbons in the feed. Similarly, steam introduced to the riser in the past has not resulted in significant reforming of the hydrocarbons in either the feed or the cracked product. In fact, the prior art suggests to the contrary that significant steam reforming reactions require high pressures (more than 100 psia), low temperatures (less than 900.degree. F.), low coke depositions (less than 1.0 Conradson carbon), and long contact times between the reaction mixture and a catalyst (more than 10 second).
The unusually large amount of coke which deposits on the catalyst in carbo-metallic oil processing presents critical problems, one problem arising from the fact that the reactions in the regenerator which convert coke to water, carbon monoxide and carbon dioxide are highly exothermic. Using a carbo-metallic feed with its unusually high content of coke precursors as compared to FCC feeds, can increase the amount of coke burned in the regenerator and the temperature in the regenerator to the point that regeneration temperatures become excessive if there is thorough burning of coke. Excessive temperatures can permanently deactivate the catalyst and/or damage the regenerating equipment.
The heat of combustion of coke depends upon the concentration of hydrogen in the coke and the ratio of CO.sub.2 to CO in the products of combustion. Carbon produces 13,910 BTU per pound when burned to CO.sub.2 and only 3,962 BTU per pound when burned to CO. Hydrogen produces 61,485 BTU per pound when burned to H.sub.2 O. The heats of combustion of coke for three representative levels of hydrogen and four different ratios of CO.sub.2 /CO are given in the following table:
TABLE I ______________________________________ Heat of Combustion BTU/lb Coke Percent Hydrogen CO.sub.2 /CO Ratio 2 4 6 ______________________________________ 0.5 8,362 9,472 10,582 1.0 11,038 12,083 3.0 14,446 4.0 12,912 14,894 ______________________________________
These problems encountered in regenerating catalysts coated with a high concentration of coke may be exacerbated when catalysts of the zeolite or molecular sieve type are used. These catalysts are crystalline aluminosilicates made up of tetra-coordinated aluminum atoms associated through oxygen atoms with silicon atoms in the crystalline structure. They have been shown to be more adversely affected by coke in respect to loss of cracking activity, than are certain other catalysts, such as for example, the non-zeolite, silica-alumina catalysts.
U.S. patent applications Ser. Nos. 94,091 (now U.S. Pat. No. 4,299,687) and 94,227 (now U.S. Pat. No. 4,354,923), filed Nov. 14, 1979, disclose processes for the conversion of carbo-metallic oils to liquid fuel in which various regeneration techniques are employed that assist in controlling the heat load in the regeneration step, each of said applications being incorporated herein by reference. One method of controlling the heat load in the regenerator is disclosed in U.S. patent application Ser. No. 251,032 for "Addition of Water to Regenerator Air" filed Apr. 3, 1981, by George D. Myers, et al., and the disclosure of this application is herein incorporated by reference.
It is thought that the ratio of CO.sub.2 to CO may be decreased to no more than about 4 and preferably to less than about 3 in order to reduce the amount of energy released within the regenerator, while optionally providing a flue gas high enough in CO content to be a useful fuel. The CO/CO.sub.2 ratio may be increased by providing chlorine in an oxidizing atmosphere within the regenerator the concentration of chlorine preferably being in the range of about 100 to about 400 ppm. This method of increasing the CO/CO.sub.2 ratio is disclosed in copending applications Ser. No. 246,751 (now U.S. Pat. 4,376,696) filed Mar. 23, 1981, for "Addition of MgCl.sub.2 to Catalyst" and Ser. No. 246,782 (now U.S. Pat. No. 4,375,404) filed Mar. 23, 1981, for "Addition of Chlorine to Regenerator", both in the name of George D. Myers. The contents of these applications are herein incorporated by reference.
As will be appreciated the carbo-metallic oils can vary widely in their Conradson carbon content. Such varying content of carbon residue in the feedstock, along with variations in riser operating conditions such as catalyst to oil ratio and others, can result in wide variations of the percent coke found on the spent catalyst. Accordingly, where the feed and riser operating conditions are such as to produce rather large coke yields, necessitating the burning of very substantial amounts of coke from the catalyst in regeneration, such as at least about 0.5 weight percent based on the catalyst, or more, chlorine additives may prove useful in controlling the heat load in the regenerator.